Saturday, 20 April 2013

WATSURF 2013: Day 5

Today we woke up to this!  Global warming is too cruel!

In any case, it was bound to be a slower day, and I took the time to work a bit.  However, I did take notes on one talk:

Branka Ladanyi (Colorado State, USA)  "Liquid confined confined in silica nanopores".  It's a simulation study on MCM-41.  Pores of diameters 20 - 40 A.  Calculated properties that can be compared with experiment: QENS (Quasi-Elastic Neutron Scattering) and OKE (Optical Kerr Effect).  Pores in simulation prepared as in Gulmen & Thompson, Langmuir 25, 1103 (2009), with OH groups on walls at density of 2 - 2.5 per nm^2.  Studied using 2-box Gibbs ensemble MC simulations.  Using the filled pore density obtained in Gibbs ensemble, perform NVT simulations of single filled pores at that density.  Simulation box is 60x60x40 A and diameters of 20, 30 and 40 A.  Using SPC/E water.  Results are more or less what you would expect:

* Density profiles show density in interior 90% of bulk.  Since pores are rough, oscillations near wall is washed out.  Slight density enhancement near wall is still visible.  Consistent with A. Soper.

* Hydrogen-bond density (n_HB / A^3) is a bit below bulk water (about 0.11 vs 0.12).  Slight peak in hydrogen bond density near wall due to water-silica H-bonds

* MSD along cylinder axis is linear with time (perpedicular to cylinder axis, trivially saturates).  You can model analytically the diffusion in confinement in a cylinder.  At long time scales, can fit diffusion constants.  In bulk, D = 2.49 x 10^-9 m^2 / s, in confinement, D ~ 1.55 for R = 10, 1.80 for R = 15 and ~2.2 (?) for R = 20 A. [ask about finite size effects on diffusion].

* If you split pore into core, surface and outer layers, find no preferential orientation in core, but orientation correlates to surface normal near surface.

* If you look at diffusion only inside core, only inside surface, find different diffusion constants (faster diffusion in core).

* Non-exponential decay in orientational correlation function C_1(t) = <P_1(u(t) . u(0))>, slower for narrower pores.  If you look only in the core region, relaxation is much faster and exponential, and equal for different diameter tubes.

* Laage & Thompson (JCP, 2012) find power-law in C_2(t) at long times for water in hydrophilic silica pores.

She then went on to map out how these observations are reflected in the experimentally-measured self-intermediate scattering function, F_s(Q, t).

With the Optical Kerr Effect, can measure polarizability anisotropy time-correlation: Psi(t) ~ < Pi^xz(0) Pi^xz(t)>, where Pi = Pi^Molecular + Pi^Induced, Pi^Molecular is just a sum of the polarizabity tensors alpha of individual molecules, and Pi^Induced is an infinite sum whose first term is something like alpha_i . T(r_ij) . alpha_j, where T(r) is the dipole-dipole interaction tensor.  I guess what's going on here is that under an applied electric field E, the dipole moment of the system is M = Pi . E.  God knows how the experiment measures <Pi^xz(0) Pi^xz(t)>.

In the afternoon, I had an interesting conversation with Werner Kuhs on how dewetting might play a role in the surface structure of methane hydrates: maybe something interesting will come of it.

No comments:

Post a Comment